Abstract A067: Precision medicine-oriented analysis of RTK-RAS signaling in gemcitabine-treated pancreatic cancer using conversational artificial intelligence

Abstract

Although KRAS mutations represent the hallmark genomic alteration in pancreatic ductal adenocarcinoma (PDAC), the broader receptor tyrosine kinase RTK-RAS signaling landscapes underlying differential responses to gemcitabine remain poorly defined. We performed an integrative clinical-genomic analysis of 150 PDAC patients stratified by age at diagnosis and gemcitabine exposure to characterize pathway- and gene-level alterations associated with treatment response. Cohort construction, pathway interrogation, and multidimensional analyses were performed using our AI-HOPE-Pancreas, a conversational artificial intelligence agent developed for precision oncology, with all principal findings independently validated using conventional statistical methods. Pathway-level alteration frequencies of RTK-RAS signaling were comparable across age- and treatment-defined subgroups, indicating that global pathway prevalence remained largely stable despite clinical stratification. In contrast, gene-level analyses uncovered distinct molecular architectures associated with both age and gemcitabine exposure. Late-onset gemcitabine-treated tumors demonstrated significant enrichment of ERBB2 and RET alterations. Conversely, TP53 mutations were more prevalent in gemcitabine-treated early-onset disease. Survival analyses further demonstrated that late-onset patients who did not receive gemcitabine and whose tumors lacked RTK-RAS pathway alterations experienced significantly improved overall survival. AI-HOPE-Pancreas enabled rapid clinical cohort generation, pathway-centric interrogation, survival modeling, and multidimensional genomic analyses while producing results consistent with conventional statistical validation. RTK-RAS signaling in PDAC extends beyond canonical KRAS mutations and exhibits distinct age- and treatment-dependent molecular architectures with potential prognostic and therapeutic relevance. These findings support pathway-informed patient stratification for precision oncology and demonstrate the utility of conversational artificial intelligence as a scalable platform for integrating multidimensional clinical and genomic data to accelerate hypothesis generation and biomarker discovery in pancreatic cancer. Camila Cosme, Brigette Waldrup, Francisco G. Carranza, Sophia Manjarrez, Vincent Chung, Laleh Melstrom, Steven Rosen, Enrique Velazquez-Villarreal. Precision medicine-oriented analysis of RTK-RAS signaling in gemcitabine-treated pancreatic cancer using conversational artificial intelligence [abstract]. In: Proceedings of the AACR Conference on Pancreatic Cancer: New Frontiers in Biology and Therapeutic Development; 2026 Sep 25-28; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(18_Suppl_2):Abstract nr A067.